MIL-101(Cr) was synthesized using a modified hydrothermal method reported by Férey (Férey et al., 2005). Typically, 332 mg of terephthalic acid (BDC) and 800 mg of Cr(NO₃)₃·9H₂O were suspended in 14 ml of H₂O, followed by the addition of HF (0.1 ml) (37%). The mixture was heated at 220°C for 8 h in a high-pressure autoclave and then cooled to 25°C to obtain a green powder. The product was washed with DMF and ethanol at 80°C for 30 min. This procedure was repeated at least three times to remove unreacted terephthalic acid and DMF. Subsequently, the green powder was activated at 150 °C for 12 h under a vacuum for further use.

Activated MIL-101 (100 mg) was suspended in 40 ml of dry n-hexane, a hydrophobic solvent, and the mixture was sonicated for 2 min. Then 0.1 ml CuCl₂ solution (0.25 M) and 0.1 ml NaOH solution (0.5 M) were added dropwise into the mixture at a rate of 5 μl/min under continuous stirring. N-hexane was carefully removed by centrifugation, and the green powder was dried at room temperature. Finally, the products were obtained by heating the green powder at 100°C for 12 h.

The morphologies of MIL-101 and CuO_x@MIL-101 were observed using scanning electron microscopy (SEM, FEI Verios-460), and elemental mapping of a bright-field scanning SEM (OXFORD Ultim Extreme, X-Max 80, Abingdon, United Kingdom) was used to confirm the composition of CuO_x@MIL-101. The crystallinity of the MOFs was determined using X-ray diffraction (XRD) (Rigaku SmartLab diffractometer with filtered Cu Kα radiation [λ = 1.5405 Å]). Moreover, the pore sizes of MIL-101 and CuO_x@MIL-101 were determined using N₂ adsorption experiments (Ipore 400 instrument [PhysiChem Instruments Ltd. Beijing, China]). Additionally, the ROS production of CuO_x@MIL-101was measured. Subsequently, CuO_x@MIL-101 was mixed with an aqueous solution of terephthalic acid (PTA) and reacted at room temperature in the dark. 300 W Xe lamp (Beijing Perfect Light) and light filter (λ > 420 nm) were irradiated at the top of the solution. The photoruminescence (PL) spectra of the solution was detected every 15 min by the fluorescence spectrophotometer (Hitachi F-4600 Model F-4600 FL Spectrophotometer) under excitation with UV light at 315 nm.

This study was approved by the Medical Ethics Committee, Zhongnan Hospital of Wuhan University (Ethics No. 2021115). Scar tissue specimens were obtained from five surgical patients, and informed consent was confirmed by the patients for experimental use. Cells from passages three to six were used for subsequent experiments.

To access the migration of HSFs treated with different concentrations of CuO_x@MIL-101 (1, 10, 100,1000 μg/ml), the scratch-wound assay was perfomed. The Cell Counting Kit 8 (CCK-8) assay (CK04; Dojindo, Kumamoto, Japan) was used to assess the viability of HSFs after treatment with these concentrations of CuO_x@MIL-101 for 24 h, 48 h, and 72 h. A hemolysis test was also performed to evaluate the safety of CuO_x@MIL-101 before its use in vivo. Detailed experimental methods were in the Supplementary Methods. This study was performed with the approval of the Experimental Animal Welfare Ethic Committee, Zhongnan Hospital of Wuhan University, Wuhan University (Approval No. ZN2022123).

Cytotoxicity assay was performed using the Cell Counting Kit 8 (CCK-8) assay (CK04; Dojindo, Kumamoto, Japan). Briefly, HSFs were seeded in 96-well culture plates (5 × 10³ cells/well) and allowed to incubate for 24 h. Cells were treated according to grouping (A:3 μg/ml MOFs +10 μg/ml CQ; B:3 μg/ml MOFs; C:10 μg/ml CQ; D: blank control group) and cultured for 12 h, then groups A and B were irradiated using a Xe lamp (Beijing Perfect Light) (30–50 mW/cm²) for 5 min. The cells were incubated for 24 h. A CCK-8 assay was performed. Cell proliferation was determined by measuring the absorbance of the different treatment groups using a microplate reader (Thermo Fisher Scientific, Waltham, MA, United States of America).

Intracellular ROS was detected by oxidation sensitive dihydroethidium (DHE, Beyotime Institute of Biotechnology, Jiangsu) fluorescent probe. The HSFs were cultured in six-well plates and treat with MOFs + Light and MOFs + CQ + Light. Following treatment, cells were harvested and co-cultured with 10 mM DHE at 37°C for 20 min according to instructions. With microplate reader, the intracellular ROS was measured. Fluorescence microscope (Zeiss, Jena, Germany) and flow cytometer (Beckman Coulter, Inc.) were used to observe intracellular ROS levels.

The mRNA expression of TGF-β1 and Collagen I in fibroblasts was detected using RT-PCR. Fibroblasts were cultured according to the grouping conditions described in section 2.3.3. Cells were washed twice with PBS, total RNA was extracted using Trizol reagent, and the RNA concentration was measured by spectrophotometry.

The western blotting assay was carried out to detect expression levels of autophagy-related proteins. The total protein of cells in the different treatment groups was extracted, separated using dodecylsulfo-polyacrylamide gel electrophoresis, transferred to polyvinylidene fluoride membranes, and mixed with anti-LC3I, LC3II, and P62 at 25°C for 1 h. The gray value of total protein was normalized to that of GAPDH, and the western blot bands were densitometrically analyzed using ImageJ software.

Microneedle patch (MNP) was prepared using a two-step pouring method as our previous study (Huang et al., 2022a). Gelatin and starch were mixed in a ratio of 2:1 and heated at 90°C. Different drugs (group A, 800 μg/ml MOFs; group B, 800 μg/ml MOFs +1.5 μg/ml CQ) were added to the mother liquor. The mixture was evenly poured into a mold. Each pour was centrifuged to remove bubbles (3000–10000 rpm) and then concentrated at 28°C–35°C for 2–3 h. This process was repeated thrice until the mold was filled with a gel. Then, the mixture without MOFs and CQ was pipetted on the surface of the MNP mold, and the substrate was prepared by centrifugation (3000–10000 rpm) in the same protocol (Figure 3A). The molded MNP was carefully removed after drying overnight for 12 h. A scanning electron microscope (SEM) was used to detect the morphology of the MNP. Young’s modulus was measured to test the mechanical strength of the MNP. To further elucidate mechanical strength and transdermal delivery capability of MNP, MNP were punctured on rabbit ear scars for 10 min. They were then fixed in 4% paraformaldehyde for 18 h to further evaluate the mechanical properties of MNP. Samples were embedded in oct compound (SAKURA Tissue-Tek^®, Torrance, CA, United States of America) and cut into 10 μm sections perpendicular to the skin surface using a microtome (Leica RM 2245, Wetzlar, Germany).

To further evaluate the efficacy of photodynamic therapy combined with microneedle therapy on scars, we established a rabbit ear scar model. All procedures for experimental animals were approved by the Experimental Animal Welfare Ethic Committee, Zhongnan Hospital of Wuhan University (Approval No. ZN2022123).

12 Japanese white rabbits weighing 2500–3000 g were chosen for this study. Sodium pentobarbital (1%, 3 ml/kg) was injected into the ear vein for anesthesia. Four squares of 1 × 1 cm size, 4 cm from the ear tip, symmetrically distributed in pairs were marked. The area was pressed to stop bleeding and wrapped with gauze. Dressings were changed every 2 days after modeling, scabs were gently removed on day 7, and modeling was completed on day 21.

The treatment efficiency for hypertrophic scar by of MNP was explored in the rabbit HS model. HSs were randomly divided into five different treatment groups. The blank control group was covered with gauze. In the MNP treatment group, the patch was pressed with the thumb and fixed with adhesive tape to immobilize (Supplementary Figure S1). Each group contained five HSs. All groups were given corresponding treatments (blank control, MOFs MNP, MOFs + light, MOFs MNP + Light, and MOFs MNP + CQ + Light) on the 1st, 8th, and 15th day after modeling. HS photographs were taken before and 1 week after each administration. Photographs were taken using a camera phone (Mi 10 Pro) to access the rabbit ear scar model during the treatment.

At the end of the treatment, the rabbits were sacrificed and ear scar specimens were collected and fixed with 4% (v/v) paraformaldehyde. All specimens were stained with H&E and Masson’s trichrome. Epithelialized collagen regeneration in the sample tissue was observed using inverted and upright fluorescence microscopy (Olympus BX51).

The scar hyperplasia index (SEI) value was calculated to quantify scar formation. The SEI was calculated as the mean of the five HSs in each group. Scarring was assessed as having an SEI value >1.5. Collagen deposition levels in HSs were assessed using collagen volume fraction (CVF). CVF (the ratio of collagen area to the total area in Masson-stained images) was calculated using ImageJ software, and five HSs were used for CVF analysis in each treatment group.

In this section, the biocompatibility of CuO_x@MIL-101 was verified using different dimensions. As shown in Figure 2A, HSFs was treated with different concentrations of CuOx@MIL-101 to evaluate its migration ability. Due to the rapid migration of HSFs in the control group (0 g/ml CuO_x@MIL-101), the scratch healed rapidly in the control group (Figure 2A, B). By comparison, HSFs migration was obviously inhibited in the 10 μg/ml CuO_x@MIL-101 group. As shown in Figure 2D, the cytotoxicity was found to be dependent on concentration, with the optimal range being 1–10 μg/ml. Therefore, 3 μg/ml of CuOx@MIL-101 was selected for further experimentation. Different concentrations of CuO_x@MIL-101 were then incubated with red blood cells at 37°C for 8 h. In Figure 2C, the supernatant of the experimental group was transparent in comparison with that of the positive control group, indicating that CuO_x@MIL-101 had almost no hemolysis occurred. Taken together, the CuO_x@MIL-101 used in this study had good biocompatibility.

In the initial stage, we identified the optimal concentration of CQ for the experiment to be between 2.5 and 10 μg/ml (Supplementary Figure S2). To achieve maximum efficiency in inhibiting autophagy, we selected 10 μg/ml as the concentration of CQ for subsequent experiments. Additionally, the optimal duration of light therapy was determined to be 5 min (Supplementary Figure S3). Additionally, we compared the therapeutic effects of two other autophagy inhibitors (HCQ and 3-MA) on HSFs. The HSFs proliferation inhibition efficacy was more stable with increasing concentrations of CQ (Supplementary Figure S4). Typically, CQ is able to inhibit autophagy by binding to lysosomes and neutralizing their acidity. This prevents the fusion of autophagosomes with lysosomes and the subsequent degradation of autophagic cargo, leading to the accumulation of autophagosomes and a decrease in autophagic flux (Yao et al., 2022). According to Figure 2E, combined with the control group, the PDT combined with CQ treatment group had the lowest cell survival rate (63% of the control group), while the PDT group had a relatively higher cell survival rate, probably because PDT could initiate protective autophagy of cells. A statistical analysis revealed a significant difference between the experimental group treated with PDT + CQ and the PDT group and the CQ group (p < 0.05), indicating that the addition of the autophagy inhibitor CQ in the photodynamic therapy became the sharp edge of PDT, thereby improving its efficacy. Furthermore, the inhibition results of the CQ + L, MOF + L, and MOF + CQ + L groups on cell proliferation align with our expectations. This suggests that CQ effectively inhibits autophagy. We can therefore conclude that the drug toxicity of CQ did not impact the experimental results.

Inhibition of inflammation and collagen hyperplasia in scar tissue is an important means of treating scars (Berman et al., 2017). A test of the expression of genes related to TGF-1 and collagen I was carried out in scar fibroblasts treated according to the group’s guidelines. As shown in Figure 2F–G, the RNA expression of Collagen I and TGF-β in scar fibroblasts of the PDT combined with the chloroquine group and the PDT group was significantly downregulated, indicating that the two treatment methods inhibited the production of type I collagen, especially in the combined treatment group.

HS is a fibroproliferative disease caused by the excessive proliferation of fibroblasts and ECM during wound healing (Zhang et al., 2020). Previous studies have demonstrated that transforming growth factor-β1 (TGF-β1) has a fibrogenic effect in hypertrophic scar fibrosis and plays a crucial part in fibroblast differentiation by regulating the expression of α-smooth muscle actin (α-SMA) in fibroblasts (Lu et al., 2005; Qian et al., 2022). α-SMA is a typical marker of the fibroblast contraction phenotype. Myofibroblasts rich in α-SMA are an important pathological basis for HS (Arora et al., 1999). Activated by TGF-β1, α-SMA-positive myofibroblasts can synthesize and secrete ECM, which further aggravates HS fibrosis (Aarabi et al., 2007). In scar tissue, AMPK has been found to be activated and it can upregulate collagen production in fibroblasts. Activation of AMPK also leads to the phosphorylation and activation of downstream targets such as p38 and Akt (Zhang et al., 2020). These downstream targets have been found to promote fibroblast activation and differentiation into myofibroblasts which are responsible for the excessive collagen production and extracellular matrix deposition in scars (Zhang et al., 2020). As expected, the expression of α-SMA was evaluated when the HSFs were administered with MOF + light irradiation, and it was found that this expression can be reversed by the presence of CQ (Supplementary Figure S5). These results strongly suggest that CQ-assist PDT can effectively decrease collagen synthesis in hypertrophic scars (HS).

In recent years, PDT has become an effective means of treating hypertrophic scars (Zhang et al., 2019; Chen Y. et al., 2021). However, studies have shown that PDT induces autophagy in a dose-dependent manner and insufficient ROS production during PDT can induce protective autophagy in HSFs (Huang et al., 2022b). LC3II/I, Beclin 1 and P62 are classical autophagy-related proteins. The maturation of autophagosomes is marked by the conversion of LC3-I to LC3-II (Pozuelo-Rubio, 2011a; Pozuelo-Rubio, 2011b; Cai et al., 2018). Therefore, LC3-II/I ratio can be used to assess autophagy. As a specific substrate of autophagy, p62 is degraded during autophagy; therefore, the intracellular level of p62 acts as a marker of autophagic flux (Klionsky et al., 2008; Rezabakhsh et al., 2018). Beclin 1 is a protein that plays a vital role in the initiation of autophagosome formation. It functions as a scaffold protein, assembling the necessary proteins for autophagosome formation and is essential for the formation of the initial pre-autophagosomal structure. The activity of Beclin 1 is regulated by various post-translational modifications and is crucial for the formation of autophagosomes. Therefore, the level of Beclin 1 protein or its activity can be used as a marker to indicate the level of autophagy in the cell (Yue et al., 2003; Fu et al., 2013).

Chloroquine (CQ), a classic autophagy inhibitor, was introduced into the treatment group to interfere with the protective autophagy of cells by preventing autophagosomes from fusing with lysosomes, resulting in the accumulation of mature, ineffective autophagic vacuoles (Redmann et al., 2017; Mauthe et al., 2018). In each group of cells, Beclin 1, LC3II/I and P62 expression was detected by western blotting (Figure 2J). An increase in the conversion of Beclin 1 was observed in the group treated with MOFs + Light + CQ, along with a increase in P62 levels. However, this increase in Beclin 1 conversion and P62 levels was reversed by the addition of CQ treatment. Despite this reversal, the level of LC3II/I was increased in the MOFs + Light + CQ group (Supplementary Figure S5). This may be due to the fact that CQ increases the acidification of lysosomes, leading to an accumulation of LC3II/LC3I in these organelles (Manic et al., 2014). This confirms that PDTcan cause cell autophagy in HSFs and that CQ can effectively block autophagosomes from fusing with lysosomes. Together, these results demonstrate that PDT induces protective autophagy in cells, thereby enhancing its anti-scarring effect.

Previous research has demonstrated that Photodynamic Therapy (PDT) can effectively increase the production of ROS in cells, leading to apoptosis (Miao et al., 2021; Zhao et al., 2021). As shown in Supplementary Figure S6, apoptosis was reduced by 5-ALA mediated PDT for HSFs when its concentration was below 4 mM. According to the mechanism of PDT, apoptosis is mostly caused by the generation of cytotoxic ROS (Li et al., 2006), and an autophagy inhibitor can block the protective autophagy induced by PDT and reduce ROS consumption (Han et al., 2017; Mohammadi et al., 2017). Therefore, ROS production of CuO_x@MIL-101 was evaluated by flow cytometry. As expected, under light irradiation, showed that ROS production was enhanced when CuO_x@MIL-101 + CQ was co-cultured with HSFs under light irradiation (Figure 2I). The separate images and quantitative analysis were in the Supplementary Figure S7. The Similar results were illustrated in the view of DHE probe (Figure 2H). Overall, the results argue that CuO_x@MIL-101 catalyzed the production of intracellular ROS following visible-light irradiation. Furthermore, the addition of CQ would boost ROS generation by inhibiting pro-survival autophagy of PDT, and resulting in optimized anti-scarring effect.

To investigate the efficacy of MNP in scar PDT treatment, we established a rabbit ear scar model. Figure 4 shows the gross images immediately after modeling. After 21 days post-surgery (the total time to establish the HS model), the treatment effect of CuO_x@MIL-101/CQ-loaded MNP on rabbits HSs was investigated using five treatment groups: MOFs MNP, MOFs + light, MOFs MNP + light, MOFs MNP + CQ + light, and a blank treatment group. After three administrations, the ear scars in the control group were raised and darker in color, while the rabbit ear scar in the MOFs MNP + CQ + light group was significantly flatter than that in the other treatment groups. MOFs MNP and MOFs + light did not significantly differ from the control group in their effects, indicating that simple MOFs smearing on rabbit ear scars for PDT and MOFs MNP treatment cannot eliminate the scars. The scar in the MOFs MNP + light group was slightly improved after treatment, but it remained dark red and had little effect. Among all the treatment groups, the rabbit ear scar after treatment in the MOFs MNP + CQ + light group was closest to the normal skin.

HS is a skin fibroproliferative disease characterized by excessive proliferation of dermal fibroblasts, excessive deposition of dermal collagen fibers, extracellular matrix protein deposition, and epithelial-mesenchymal transition (EMT) (Zhang et al., 2016; Lee and Jang, 2018). We evaluated the status of HS using microscopic observation of histopathological features in the different treatment groups. The scars were stained with HE and Masson’s trichrome to observe histological changes and quantitatively evaluate SEI and CVF.

Figure 5A shows the results of HE staining. The number of fibroblasts in the MOFs MNP + CQ + light group is the least, and the other groups have different degrees of neutrophil infiltration. The SEI results for each group are shown in Figure 5C. As predicted, control group SEI was significantly higher, and MOFs MNP and MOFs + light results were not significantly different from control group SEI. Consistent with the results presented in the general photographs of rabbit ear scars after treatment, the SEI of the MOFs MNP + light group was lower than that of the PDT group treated with MOFs alone, but the difference was not significant. However, the SEI of the MOFs MNP + CQ + light group was significantly reduced, indicating that the MOFs was introduced into the deep scar tissue by microneedles, and the MOFs (CuO_x@MIL-101) based PDT indicated a good an inhibitory effect on the formation and growth of hypertrophic scars.

Additionally, Masson staining showed (Figure 5B) that compared with the dense, curly, and irregular collagen arrangement in the blank control group, the collagen arrangement in the scar tissue of the MOFs MNP + CQ + light group was looser and more regular, and the direction was consistent with the epidermis. Figure 5D shows the quantitative measurement of collagen volume, and the results also show that MOFs MNP combined with PDT treatment can significantly improve collagen deposition in scar tissue. Notably, the best inhibitory effect was realized in the combined treatment by MOF-assisted PDT and CQ. As a result, the collagen deposition process was blocked by the MOF-assisted PDT and its anti-scarring efficacy can be boosted by CQ.